Coding
Like-for-like- Gemini 3 Flash
- 41.5
- Supported · #108/151
- GPT-5.4 mini
- 42.7
- Supported · #104/151
- Basis
- BenchAlign lane · 3 vs 4 public rows
- Reading
- GPT-5.4 mini leads · intervals overlap
Keep up with the models you depend on. Follow price changes, retirements, and API updates.Follow the models you depend on.
Follow model changesUpdated September 10, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
Gemini 3 Flash has the higher public score estimate, 62.2 versus 61.11, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
8 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
Share or export
Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.
Code generation, repair, and software-engineering tasks
GPT-5.4 mini
GPT-5.4 mini leads on the public coding lane, 42.7 to 41.5, with Supported evidence for both models, although the 90% intervals overlap.
Confidence: limited
Prompts that approach the documented context limit
Gemini 3 Flash
Gemini 3 Flash has the larger documented context window.
Confidence: documented
1K fresh input + 500 output tokens
Gemini 3 Flash
Gemini 3 Flash has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
200K cached + 20K fresh input + 10K output tokens
Gemini 3 Flash
Gemini 3 Flash has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
50K fresh input + 3K output tokens
Gemini 3 Flash
Gemini 3 Flash has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Tool use, computer use, and multi-step task completion
Not enough matched evidence
GPT-5.4 mini is scored on Estimated evidence for agentic, so the reading is directional rather than like-for-like.
Confidence: limited
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
2 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Each row shows the public-lane category score for both models: the BenchAlign lane for agentic, coding, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.
| Category | Gemini 3 Flash | GPT-5.4 mini | Basis | Reading |
|---|---|---|---|---|
| Coding | 41.5Supported · #108/151 | 42.7Supported · #104/151 | Like-for-likeBenchAlign lane · 3 vs 4 public rows | GPT-5.4 mini leads · intervals overlap |
| Knowledge | 56.9Supported · #45/183 | 55.6Supported · #52/183 | Like-for-likeBenchAlign lane · 2 vs 5 public rows | Gemini 3 Flash leads · intervals overlap |
| Agentic | 34.0Supported · #134/152 | 39.1Estimated · #119/152 | Directional onlyBenchAlign lane · 4 vs 6 public rows | Directional only |
| Instruction following | 66.2#68/123 | 89.8#23/123 | Directional onlyProvisional lane · 0 vs 0 weighted rows | Directional only |
| Reasoning | 59.0Unranked · 2 rankable rows | 73.9Unranked · 2 rankable rows | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | 50.2Unranked · 2 rankable rows | 44.5Unranked · 2 rankable rows | Not comparableProvisional lane · 2 vs 2 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multimodal | 75.8Unranked · 1 rankable row | 57.2#31/48 | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign) differ from the provisional-lane categories. Unranked scores sit on the lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.
FrontierMath v2 (Tiers 1-3)
Math
LiveCodeBench (Vals)
Coding
MMLU-Pro (Vals)
Knowledge
FrontierMath v2 (Tier 4)
Math
Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.
1K fresh input + 500 output tokens
Gemini 3 Flash has the lower modeled cost
Costs use the listed standard API rates.
50K fresh input + 3K output tokens
Gemini 3 Flash has the lower modeled cost
Costs use the listed standard API rates.
200K cached + 20K fresh input + 10K output tokens
Gemini 3 Flash has the lower modeled cost
Costs use the listed standard API rates.
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
Maximum documented context; output-token limits may be lower.
Gemini 3 Flash
GPT-5.4 mini
Gemini 3 Flash
gemini-3-flash-preview
Google Gemini API pricingGPT-5.4 mini
gpt-5.4-mini
OpenAI GPT-5.4 mini model documentationA missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Gemini 3 Flash
$0.05 per 1M cached input tokens
Google Gemini API pricingGPT-5.4 mini
$0.075 per 1M cached input tokens
OpenAI pricingGemini 3 Flash
Not sourced
GPT-5.4 mini
text, image
OpenAI model catalogGemini 3 Flash
Not sourced
GPT-5.4 mini
Gemini 3 Flash
Not sourced
GPT-5.4 mini
Generally Available · OpenAI Responses API
OpenAI model catalogGemini 3 Flash
Non-Reasoning
GPT-5.4 mini
Reasoning
Gemini 3 Flash
Proprietary
GPT-5.4 mini
Proprietary
Gemini 3 Flash
Proprietary
GPT-5.4 mini
Proprietary
Gemini 3 Flash
2025-12-01
GPT-5.4 mini
2026-03-17
Run the same representative tasks against both endpoints before changing production traffic.
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Claw-Eval
Not directly comparable
Gert Labs
Not directly comparable
JobBench
Not directly comparable
Terminal-Bench 2.1 (Vals)
GPT-5.4 mini leads this result
Terminal-Bench 2.0
Not directly comparable
OSWorld-Verified
Not directly comparable
MCP Atlas
Not directly comparable
Toolathlon
Not directly comparable
τ²-bench results
Not directly comparable
Vibe Code Bench
Shared sourceGPT-5.4 mini leads this result
LiveCodeBench (Vals)
Gemini 3 Flash leads this result
SWE-bench (Vals)
Gemini 3 Flash leads this result
FrontierCode 1.1 Main
Not directly comparable
GPQA Diamond (Vals)
Gemini 3 Flash leads this result
MMLU-Pro (Vals)
Gemini 3 Flash leads this result
GPQA
Not directly comparable
HLE
Not directly comparable
HLE w/o tools
Not directly comparable
FrontierMath v2 (Tiers 1-3)
Shared sourceGemini 3 Flash leads this result
FrontierMath v2 (Tier 4)
Shared sourceGemini 3 Flash leads this result
Gemini 3 Flash has the higher public score estimate, 62.2 versus 61.11, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
GPT-5.4 mini leads the public coding lane, 42.7 to 41.5, with Supported evidence for both models, although the 90% intervals overlap.
GPT-5.4 mini scores higher for agentic tasks on the public lane, 39.1 to 34. GPT-5.4 mini is scored on Estimated evidence for agentic tasks, so the reading is directional rather than like-for-like. BenchLM does not name a winner for agentic tasks from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
For the stated presets, chat costs $0.002 on Gemini 3 Flash and $0.003 on GPT-5.4 mini; repository review costs $0.034 and $0.051; the cache-heavy agent loop costs $0.05 and $0.075. Costs use the listed standard API rates.
Gemini 3 Flash has the larger documented context window: 1M, compared with 400K.
Last updated September 10, 2026
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.